Integrated Analysis of the Effects of Cecal Microbiota and Serum Metabolome on Market Weights of Chinese Native Chickens

Simple Summary Native chickens generally have the characteristics of low growth performance, which has also become a limiting factor for the breeding of yellow feather broilers. More and more studies have shown that gut microbiota plays an important role in the growth of livestock farming. The Guizhou yellow chicken is a breed of yellow-feathered broiler chicken with excellent meat quality and good flavor currently being cultivated in Guizhou Province, China. In order to explore the role of gut microbiota on the growth performance of native chickens, the Guizhou yellow chicken was taken as a representative, and high-market-weight and low-market-weight chicken groups were established according to their market weights. By integrating microbial 16S rRNA gene sequencing and non-targeted serum metabolome data, five key cecal microbes associated with high body weight in chickens and one key microbe associated with low body weight were identified. In addition, the results also showed that specific gut microbes might positively affect the growth rate of chickens by regulating vitamin and other metabolic pathways. These findings might improve understanding of the role of gut microbiota in chicken growth traits and their underlying metabolic mechanisms. Abstract The gut microbiota plays an important role in the physiological activities of the host and affects the formation of important economic traits in livestock farming. The effects of cecal microbiota on chicken weights were investigated using the Guizhou yellow chicken as a model. Experimental cohorts from chickens with high- (HC, n = 16) and low-market-weights (LC, n = 16) were collected. Microbial 16S rRNA gene sequencing and non-targeted serum metabolome data were integrated to explore the effect and metabolic mechanism of cecal microbiota on market weight. The genera Lachnoclostridium, Alistipes, Negativibacillus, Sellimonas, and Ruminococcus torques were enriched in the HC group, while Phascolarctobacterium was enriched in the LC group (p < 0.05). Metabolomic analysis determined that pantothenic acid (vitamin B5), luvangetin (2H-1-benzopyran-6-acrylic acid), and menadione (vitamin K3) were significantly higher in HC serum, while beclomethasone dipropionate (a glucocorticoid) and chlorophene (2-benzyl-4-chlorophenol) were present at higher levels in the LC group. The microbes enriched in HC were significantly positively correlated with metabolites, including pantothenic acid and menadione, and negatively correlated with beclomethasone dipropionate and chlorophene. These results indicated that specific cecal bacteria in Guizhou yellow chickens alter the host metabolism and growth performance. This study provides a reference for revealing the mechanism of cecal microbe actions that affect chicken body weight.


Introduction
Chicken meat is a major protein source throughout the world [1], and the improvement of chicken production performance is necessary to meet increasing demand.Market weight is a key growth trait in chickens, and high market weight can increase the turnover in the chicken production pen and reduce labor costs for farmers [2].Several factors, such as genetics [3], management [4], and nutrition [5], can affect market weight, thus it can be scientifically altered.Gut microbes might be involved in the regulation of chicken market weight.For example, the presence of Lachnospiraceae in the cecum was found to be related to high growth performance (body weight) for chickens, while the presence of Escherichia had the opposite effect [6].Cecal microbiota possessing Microbacterium and Sphingomonas in Turpan gamecock progeny × White Leghorn chickens were significantly correlated with high body weight, while Slackia was enriched in the ceca of low-market-weight chickens [7].In addition, high Lactobacilli abundance in the chicken jejunum was beneficial to growth, while Comamonas enrichment produced a negative outcome on growth rates [8].The addition of exogenous Bacillus subtilis and Bacillus licheniformis to chicken feed also promoted growth performance [9,10].However, the conclusions have not been completely consistent in determining which microbes produce the greatest effect on chicken growth performance.
In spite of numerous studies focused on unraveling the mechanisms for market weight increases, this complex trait has yet to be fully understood [11][12][13].Since metabolite production might profoundly alter host physiological functions [14][15][16] and thereby affect host phenotypes, such as feed efficiency [17], disease resistance [18], meat quality [19], and gut microbiome can influence host metabolism [15], it is deduced that the influence of gut microbiota on host metabolism might be one of the mechanisms to regulate growth performance of chickens.However, few researches have been found.Only one study proved that cecal microbiota could affect chicken growth performance by regulation of fat metabolism [7].Unfortunately, there was a limitation in this study, for it focused on only a few specific lipid metabolites.As well known, there are several host metabolites, and the regulatory role of other metabolites remains unclear.Thus, more advanced techniques are necessary for a more comprehensive evaluation.
As a modern omics technology, metabolomics allows high-throughput, multi-dimensional analysis of a large number of metabolites with high sensitivity and can detect very small metabolic changes [20].Although it has not been directly employed for study on chicken growth performance, it has been applied to fill the information gap between gene and phenotype [21].With this technology, more comprehensive metabolite information can be obtained.A combination of microbiomes and metabolomics linked gut microbes to the regulation of metabolite production, which is becoming an important means of analyzing complex traits of livestock farming [22][23][24][25].Based on this, researchers have explored the mechanisms of fat deposition in pigs and milk protein formation in cows.For example, Prevotella is a key bacterial genus that affects pig intramuscular fat deposition via the production of lipopolysaccharides, branched-chain amino acids (BCAAs), and arachidonic acids [26].Prevotella abundance in the rumen altered amino acid metabolism, resulting in increased milk protein content.In contrast, the enrichment of methanogens in the rumen was not conducive to increasing the milk protein content [27].In chicken, multi-omics technology was used for the characterization of bacillus spp.probiotics isolated from European broilers to improve their growth performance [28].Therefore, integrating gut microbiome and metabolomics is helpful for the identification of key microbes related to growth traits and the analysis of their metabolic mechanisms in chickens [29,30].
Guizhou yellow chicken is a breed of yellow-feathered broiler chicken with excellent meat quality and good flavor currently being cultivated in Guizhou Province, China.This synthetic chicken breed was developed by crossing parental breeds (Guizhou Weining chicken as female parent and Golden Plymouth Rock chicken as male parent) [31].However, compared with commercial broiler breeds such as Ross and Cobb [32], the growth rate of this breed is slow [33].Furthermore, the factors that influence their growth rates have not yet been fully elucidated.
In this study, it was hypothesized that the specific cecal microbiota might be involved in regulating the market weight of Guizhou yellow chickens by influencing host metabolism.Microbial 16S rDNA gene sequencing and untargeted metabolomics were combined to explore the effect of gut microbiome on the growth performance of Guizhou yellow chickens and to identify key microbes associated with market weight and preliminary explore the possible metabolic mechanisms for improved growth performance.The results of this study could give insight into the analysis of chicken growth performance and provide research references for the discovery of growth-promoting probiotics.

Ethics Statement
The Animal Care and Use Committee at Guizhou University approved this project (approval number: EAE-GZU-2022-T050).All animal works were conducted according to the guidelines for the care and use of experimental animals established by the Ministry and Rural Affairs of the People's Republic of China.

Experimental Animals and Sample Collection
The animals involved in this study were Guizhou Yellow chickens bred in our laboratory and raised in the chicken farm of Guizhou University (research farm) from June to October 2022.All chickens (n = 49, 24 female + 25 male) were hatched on the same day and were raised in cages.Stocking density was as follows: at the age of 0-4 weeks, 4-10 weeks, and 10-18 weeks, 16 chickens (male and female), 8 chickens (male and female), and 1 chicken were kept in each cage, respectively.The ambient feeding temperature was set at 26 • C when the animals were 0-4 weeks old; after that, chickens were transferred to the roller-curtain natural ventilation cooling chicken house without a temperature regulation system and raised at natural room temperature.The temperature was 20-35 • C in June-August and 15-25 • C in September-October.All chickens were fed at 5 AM and 5 PM every day.The light duration was about 16 h per day, and the humidity was 60% to 65%.All chickens were raised in the same environment.The chicken house was cleaned and disinfected regularly according to the sanitary and epidemic prevention requirements of the farm to keep them clean, dry, and ventilated.All the chickens were allowed to eat and drink freely.At different growth stages, the animals were fed brood feed, nursery feed, and growing feed according to different nutritional requirements.The nutritional composition is shown in Table 1.The chickens were weighed at the same time on the morning of the weighing day using an electronic scale every 2 weeks (±5 g).At 18 weeks of age, 8 roosters and 8 hens with the highest body weight were selected into the high-market-weight group (HC, n = 16), and 8 roosters and 8 hens with the lowest body weight were selected into the low-market-weight group (LC, n = 16).Namely, in each group, 16 replicates were involved.No antibiotics were used within one month before slaughter.At the end of the experiment (18-week-old), serum samples and cecal contents were collected from all 32 chickens (16 in the high-market-weight group and 16 in the low-market-weight group).Whole blood was obtained from wing veins for serum separation before the chickens were euthanized by CO 2 asphyxiation [34], and then ~2 g cecal content was collected at the same position as their cecum.Serum samples and cecal content were immediately put into liquid nitrogen for quick freezing after isolation or collection and were stored at −80 • C for further analyses.Thirty-two serum samples (16 from HC, 16 from LC) for untargeted metabolomics and 32 cecal content samples (16 from HC, 16 from LC) for 16S rDNA gene sequencing were performed at Shanghai Applied Protein Technology (Aptbio, Shanghai, China) and Shanghai Majorbio Biopharm Technology (Majorbio, Shanghai, China), respectively.The experimental flow chart is shown in Figure 1. ).Cecal samples were collected and subjected to 16S rRNA sequencing to infer microbial profiles.Concurrent blood samples were collected to perform untargeted metabolomics detection.Cecal and serum metabolites were identified by statistical analysis.

Sequence Splicing and ASV Annotation
Amplicon sequencing was conducted using an Illumina MiSeq platform (Illumina, San Diego, CA, USA).The reads were first filtered and assembled into tags according to Figure 1.Experimental flow chart.The experimental cohort comprises 49 healthy Guizhou yellow chickens; 8 roosters and 8 hens that possessed the highest body weights in the group constituted the high-market-weight group (HC, n = 16), and a similar group was selected for the low-market-weight group (LC, n = 16).Cecal samples were collected and subjected to 16S rRNA sequencing to infer microbial profiles.Concurrent blood samples were collected to perform untargeted metabolomics detection.Cecal and serum metabolites were identified by statistical analysis.

Sequence Splicing and ASV Annotation
Amplicon sequencing was conducted using an Illumina MiSeq platform (Illumina, San Diego, CA, USA).The reads were first filtered and assembled into tags according to overlap relationships between the paired-end reads.Tags were then clustered into amplified sequence variants (ASVs) [35].Data was optimized using the DADA2 to obtain the representative ASV sequence and abundance information [36].Representative sequences of each ASV were annotated using the Silva database (https://www.arb-silva.de/)(accessed on 22 February 2023) with a taxonomic confidence level of 0.7 [37].

Extraction of Serum Samples
Serum samples were thawed at 4 • C, and an appropriate amount was added to precooled a methanol/acetonitrile/water solution (2:2:1, v/v) and vortexed and ultrasonicated at low temperature for 30 min and kept at −20 • C for 10 min.Samples were then centrifuged at 14,000× g at 4 • C for 20 min.Supernatants were transferred into clean tubes and dried under vacuum, and the residue was suspended in 100 µL 50% acetonitrile for UPLC-QTOF/MS analysis [38].
2.5.Chromatography-Mass Spectrometry Analysis 2.5.1.Chromatographic Conditions Sample compounds were separated using an Agilent (Agilent Technologies, Santa Clara, CA, USA) 1290 Infinity LC ultra-high performance liquid chromatography system using a hydrophilic interaction (HILIC) column using the mobile phases as follows: A (25 mM ammonium acetate and 25 mM ammonia in H 2 O) and B (acetonitrile).The gradient elution procedure is as follows: 0-0.5 min, 95% B; 0.5-7 min, B 95 to 65%; 7-8 min, B 65 to 40%; 8-9 min, B 40%; 9-9.1 min, B 40 to 95%; 9.1-12 min, B 95%.During this process, the proportion of mobile phase A changed accordingly.Samples were placed in an autosampler tray at 4 • C. QC samples were inserted into the sample queues to monitor and evaluate the stability of the system and the reliability of experimental data [39].

Data Processing
Proteo Wizard MS Convert was used to convert the raw MS data to MzXML files before importing it into freely available XCMS software (v 1.52.0)[40].For peak picking, the following parameters were used: centWave m/z = 10 ppm, peakwidth = c (10, 60), prefilter = c (10, 100).For peak grouping, bw = 5, mzwid = 0.025, minfrac = 0.5 were used.Isotopes and adducts were annotated by CAMERA (Collection of Algorithms of MEtabolite pRofile Annotation) [41].In the extracted ion features, only the variables with more than 50% of the nonzero measurement values in at least one group were kept [41].Compound identification of metabolites was conducted by comparing accuracy m/z value (<10 ppm) and MS/MS spectra with an in-house database established with available authentic standards [42].

Statistical Analysis 2.6.1. Analysis of Microbiota Diversity and Composition Differences
Analysis for taxonomic clusters utilized the ASVs, and quality control was carried out under conditions of relative abundance >0.05% and detection of ASVs in >80% of the samples [43].ASV α-diversity with Shannon, Simpson, Chao1, Faith's phylogenetic diversity (PD), and ACE indices were calculated using Mothur software v 1.31.2 [44].The Wilcoxon rank sum test was used to compare α-diversity differences between the two groups.Principal coordinates analysis (PCoA) [45] was performed to evaluate the discrepancy of the phylogenetic compositions of cecal microbiota between HC and LC.Linear discriminant analysis with difference contribution analysis (LEfSe) was performed with LDA > 2 and p < 0.05 as thresholds to identify bacterial composition differences between the two groups [46].

Construction of Cecal Microbial Co-Abundance Groups
The quality-controlled ASVs were used to construct co-abundance groups (CAGs) of cecal microbiota.The correlation matrix between ASVs was calculated based on the SparCC algorithm [47] using the SpiecEasi package in the R [48].Paired ASVs with a correlation coefficient >0.5 were used for further analyses.The correlation coefficient values were converted into correlation distances (1-correlation coefficient), and ASVs were clustered into CAGs based on the Ward algorithm with the Vegan package in the R [49].Cytoscape v 3.9.1 [50] was used for the visualization of cecal microbiota CAGs.

Serum Metabolomics Analysis
After sum-normalization, the online analysis platform MetaboAnalyst 5.0 (https: //www.metaboanalyst.ca/)(accessed on 14 March 2023) was employed to analyze the processed data by multivariate data analysis, including Pareto-scaled principal component analysis (PCA) and orthogonal partial least-squares discriminant analysis (OPLS-DA) [51].The robustness of the model was evaluated using 7-fold cross-validation and response permutation testing.The variable importance in the projection (VIP) value of each variable in the OPLS-DA model was calculated to indicate its contribution to the classification [52].In order to determine the significance of differences between two groups of independent samples, the Student's t-test was applied.VIP > 1 and p < 0.05 were used to screen significantly changed metabolites.Pearson's correlation analysis was performed to determine the correlation between two variables [49].

Growth Performance of the Study Chickens
The weight gain trends for the HC and LC groups of chickens were consistent from 0 to 18 weeks of age when the animals were provided with identical environments and feed and water access.There was no significant difference in body weights between the two groups from 0 to 6 weeks of age.In contrast, from week 8 onwards, body weight differences reached the level of statistical significance (p < 0.05) (Figure 2A).By 18 weeks of age, the body weights for HC (2187.81± 232.58 g) were significantly (p < 0.01) greater than the LC group (1817.69± 199.30 g) (Figure 2B and Table 2).

Cecal Microbial Diversity in High-versus Low-Market-Weight Chickens
A total of 32 microbial DNA samples from HC and LC chickens were used for 16S rDNA gene sequencing, 1,535,941 clean reads were generated after QC, and 1955 ASVs were identified for all samples (Table S1).The Shannon and Simpson of α-diversity indices

Cecal Microbial Diversity in High-versus Low-Market-Weight Chickens
A total of 32 microbial DNA samples from HC and LC chickens were used for 16S rDNA gene sequencing, 1,535,941 clean reads were generated after QC, and 1955 ASVs were identified for all samples (Table S1).The Shannon and Simpson of α-diversity indices between the two groups were of no significant differences (Figure 3A,B).PCoA analysis showed a lack of significant differences in β-diversity of the HC and LC cecal microbiota, although a certain aggregation effect was evident (Figure 3C).

Identification of Co-Abundance Groups (CAG) Associated with Body Weight
Another goal of this study was to explore the gut microbiota clusters associated with chicken body weight [53].The identified 223 ASVs were clustered into 20 CAGs, and the average relative abundance of each CAG was compared between groups using the Wilcoxon test.Four CAGs were identified that significantly differed between the two groups: CAG6, CAG9, CAG16, and CAG17 (Figure 5A).CAG6 was enriched in LC and contained 22 ASVs, and Phascolarctobacterium ASV5 was the most enriched in LC.This suggested that CAG6 with Phascolarctobacterium as the core has a potentially negative effect on growth performance.In contrast, CAG17 was enriched in HC and included Ruminococcus torques group ASV642, Bifidobacterium ASV279, Alistipes ASV26, Alistipes ASV239, and others.In addition, CAG9 was centered on Ruminococcus torques group ASV20, while CAG16 contained bacteria such as Parabacteroides that were enriched in HC at the genus level (Figure 5B and Table S4).
performance.In contrast, CAG17 was enriched in HC and included Ruminococcus torques group ASV642, Bifidobacterium ASV279, Alistipes ASV26, Alistipes ASV239, and others.In addition, CAG9 was centered on Ruminococcus torques group ASV20, while CAG16 contained bacteria such as Parabacteroides that were enriched in HC at the genus level (Figure 5B and Table S4).

Correlation Analysis Reveals Relationships between the Cecal Microbiota and Serum Metabolites
The overlaps of differential bacteria at the ASV, genus, and CAG levels were integrated.The Ruminococcus torques group, Lachnoclostridium, Alistipes, Negativibacillus, and Sellimonas were enriched in HC, and Phascolarctobacterium was enriched in LC in all analyses at three levels (Table S5).Spearman's rank correlation analysis was then performed between these bacteria and differential metabolites, and the Ruminococcus torques group enriched in HC was significantly positively correlated with pantothenic acid (p = 0.029, r = 0.386).In addition, menadione was positively correlated with numerous bacteria enriched in HC.These included Lachnoclostridium ASV482, Alistipes ASV26, Negativibacillus ASV83, and Sellimonas ASV409, suggesting these bacteria might promote the synthesis of menadione and have beneficial effects on the growth performance of the chickens.The detailed relationships between weight-related bacteria and differential metabolites are shown in Figure 7.

Discussion
Growth performance is an important trait for chickens, and gut microbiota have been proven to play an important role in the life activities of hosts.However, clear correlations between the gut microbes and chicken growth performance have not been fully revealed.In this study, multi-omics was used to describe the cecal microbiota composition and serum metabolite differences between high-and low-market-weight Guizhou Yellow chicken.The results showed that Lachnoclostridium, Alistipes, Negativibacillus, Sellimonas, and Ruminococcus torques are beneficial for chicken growth by regulating metabolites such as pantothenic acid and menadione, while Phascolarctobacterium might inhibit the growth of chicken.
The composition of the gut microbiota for individual animals also varies by location within the animal, and the diversity and abundance of microbiota in the cecum is the highest.Therefore, cecal microbiota was taken as representative gut microbiota in this study [54][55][56].Diversity is an important index to evaluate the community structure of microbiota [57].It has been proved that the diversity of gut microbiota is negatively correlated with weight gain [58].Decreased diversity of gut microbiota has been linked to inflammatory diseases, which tend to result in fat deposition and increased body weight [26].However, results from other studies found no significant difference between the cecal microbiota diversity of high-market-weight and low-market-weight chickens [7], which is consistent with the observation in this study.These results are controversial as of now;

Discussion
Growth performance is an important trait for chickens, and gut microbiota have been proven to play an important role in the life activities of hosts.However, clear correlations between the gut microbes and chicken growth performance have not been fully revealed.In this study, multi-omics was used to describe the cecal microbiota composition and serum metabolite differences between high-and low-market-weight Guizhou Yellow chicken.The results showed that Lachnoclostridium, Alistipes, Negativibacillus, Sellimonas, and Ruminococcus torques are beneficial for chicken growth by regulating metabolites such as pantothenic acid and menadione, while Phascolarctobacterium might inhibit the growth of chicken.
The composition of the gut microbiota for individual animals also varies by location within the animal, and the diversity and abundance of microbiota in the cecum is the highest.Therefore, cecal microbiota was taken as representative gut microbiota in this study [54][55][56].Diversity is an important index to evaluate the community structure of microbiota [57].It has been proved that the diversity of gut microbiota is negatively correlated with weight gain [58].Decreased diversity of gut microbiota has been linked to inflammatory diseases, which tend to result in fat deposition and increased body weight [26].However, results from other studies found no significant difference between the cecal microbiota diversity of high-market-weight and low-market-weight chickens [7], which is consistent with the observation in this study.These results are controversial as of now; more studies are therefore needed.
It was found in this study that at the genus level, Lachnoclostridium, Alistipes, Negativibacillus, Sellimonas, and Ruminococcus torques were enriched in the HC group, while Phascolarctobacterium was enriched in the LC group.These specific cecal microbes might be an important factor influencing the body weight of chickens.Alistipes and Lachnoclostridium are considered to be important producers of short-chain fatty acids, including butyric and acetic acids.It has been proved that a reduction in Alistipes abundance is linked to a reduction in the levels of short-chain fatty acids [59,60], and Alistipes finegoldii was specifically proved to promote the growth of broiler chickens [60][61][62].In mice, a reduced Lachnoclostridium abundance was associated with decreased body weight [63].The reason might be that in addition to influencing the production of short-chain fatty acids, Lachnoclostridium is also linked to host nutrient absorption, and its reduced abundance will lead to the downregulation of functional pathways such as protein processing and nutrient transport in the host [59,64,65].Negativibacillus is a Gram-negative Firmicute, and its abundance in the mouse gut was positively correlated with body weight gain [66].Sellimonas is an obligate anaerobic, non-motile Gram-positive first isolated from human feces in 2016 [67].In addition, Ruminococcus torques were found to be enriched in chickens with high body weight.It has not been studied much since Ruminococcus torques was first described.However, it is often involved in studies that link human microbiota to disease states.For instance, it has decreased abundance in patients with Crohn's Disease in comparison to healthy individuals [68].In contrast, its abundance increased in children with late-onset autism [69] and those with autism spectrum disorders and gastrointestinal disorders [70].Therefore, the role of Ruminococcus torques in contribution to health or disease is still an open question [71].
Previous studies have found that there are several mechanisms explaining how gut microbiota affect the growth of the host, such as being involved in vitamin synthesis [72,73], dietary fiber degradation [74], inflammatory induction [33], and lipid metabolism [7].As well known, vitamins play a catalytic role in promoting nutrient synthesis, thereby controlling metabolism and affecting the performance and health of poultry [75].Humans and animals cannot synthesize most vitamins by themselves and must obtain them from their diet or rely on gut microbiota to synthesize them [76].Probiotics such as Bifidobacterium and lactobacillus could synthesize a variety of vitamins necessary for human growth and development, such as vitamins B [72], vitamin K [77], and vitamin D [73].In this study, several vitamin metabolites related to body weight were identified by serum metabolome analyses.For example, pantothenate acid and menadione were enriched in the serum of high-market-weight chickens, while riboflavin was enriched in the serum of low-marketweight chickens, which highlights the importance of vitamin metabolism for the regulation of growth traits in chickens.
Pantothenic acid, also known as vitamin B5, is an indispensable essential nutrient that can be converted into Coenzyme A (CoA) and Acyl carrier protein (ACP) in living organisms, both of which are enzyme cofactors necessary for key pathways of metabolism and energy production in all living cells [78].Pantothenic acid is involved in the metabolism of sugar, fat, and protein in both humans and animals [79].In this study, a higher concentration of pantothenic acid was found in the serum of high-market-weight chickens.This indicates that pantothenic acid could promote the growth and development of broilers.Early studies showed that deficiency of pantothenic acid caused symptoms such as loss of appetite, growth retardation, and wasting in sick chickens [80,81].Lacking pantothenic acid in the diet for chicks could result in a decrease in protein, fat, and energy stores, and the addition of pantothenic acid in the diet could not only enhance the activity of intestinal digestive enzymes but also promote the digestion and absorption of nutrients in the diet, thus promoting the growth of animal body [82].In addition, it was found in this study that Ruminococcus torques were enriched in high-market-weight chickens and were significantly positively correlated with pantothenic acid, suggesting that Ruminococcus torques might promote the generation of pantothenic acid and, in turn, promote the growth of chickens.
In this study, a significantly higher concentration of menadione (VK 3 ) in high-marketweight chickens was found.It is deduced that the beneficial effects of menadione on bone development and its strong antioxidant effect might be two of the reasons for the chickens' higher market weight.Menadione is a fully reduced form of vitamin K.In addition to its well-known anticoagulant effects, vitamin K also plays an important role in bone formation and remodeling [75].It has been proved that supplementation of vitamin K in the diets of mg/kg feed) and grower (2 mg/kg feed) broilers promoted the carboxylation of osteocalcin and improved the hydroxyapatite binding ability of serum osteocalcin, in turn, improved the bone quality [83].What's more, menadione has a strong antioxidant effect [84].Oxidative stress adversely affects the growth performance of animals because it can cause disorders of chicken gastrointestinal peristalsis, which tend to result in enteritis and diarrhea, damage of intestinal villi, in turn, lead to poor absorption of nutrients, and reduce the nutrient absorption capacity of chickens [85].The chickens involved in this study were raised from June to October when the temperature was up to 33 • C; natural ventilation might not cool the chicken house effectively, which might cause heat stress in chickens.Some chickens with poor tolerance to heat stress might grow slower, while the high-market-weight chickens may alleviate heat stress by increasing the synthesis of menadione.
It was early discovered that menadione could be produced by many bacteria, such as Bacillus cereus, B. mycoides, B. subtilis, Chromobacterium prodigiosus, Escherichia coli, Mycobacterium tuberculosis, Sarcina lutea, and Staphylococcus aureus [86], and human gut microbiota (e.g., Bacteroides and Prevotella) might participate the synthesis of menadione [87].Interestingly, menadione was found to be enriched in high-market-weight chickens and was significantly associated with a variety of high-market-weight-associated bacteria, including Lachnoclostridium, Alistipes, Negativibacillus, and Sellimonas.This suggests that these microbiotas might promote the synthesis of menadione and thus affect the growth of chickens.However, its mechanism needs further study.
In the present study, Phascolarctobacterium was found to be a key microbe that was enriched in the cecum of low-weight chickens.Phascolarctobacterium is an obligate anaerobic originally isolated from koala feces [88].Another study also found that Phascolarctobacterium abundance was higher in low-feed conversion chickens and was related to a low nutrient absorption capacity of the host.However, the study did not elucidate a detailed mechanism [89].An additional study found that the abundance of Phascolarctobacterium in the gut increases when chickens are exposed to high temperatures for extended periods, and this exposure also results in elevated levels of heat shock proteins and related inflammatory gene expression [90].Heat stress altered the structure and function of enzymes in the chicken body, reduced the pH of the blood, and caused metabolic acidosis.These were negative influences on chicken growth [91,92], and intestinal inflammation also decreases nutrient absorption, causing body weight to drop [93][94][95].The presence of Phascolarctobacterium was correlated with the induction of inflammation under the action of heat stress and other harmful factors in chickens.These results are consistent with those of this study, where Phascolarctobacterium was enriched in the cecum of low-market-weight chickens.As mentioned above, elevated serum levels of some vitamin-related metabolites in high-market-weight chickens might contribute to the relief of heat stress.
In addition, it is worth noting that in this study, 11 beta-hydroxyprogesterone showed higher concentration in the serum of low-market-weight chickens.11 beta-hydroxyprogesterone is a naturally occurring, endogenous steroid and derivative of progesterone [96].This might be because at the age of 18 weeks, the low-market-weight chickens were already sexually mature, and the follicular granulosic cells released related sex hormones, making them lay eggs.The onset of sexual maturity of chickens might slow down their growth because the nutrients would be utilized for reproduction instead of growth [97].

Conclusions
In this study, 16S rRNA gene sequencing with untargeted serum metabolomics was combined to identify cecal microbes associated with body weight in chickens and to preliminarily identify the metabolic mechanism of cecal microbiota affecting growth performance.The effect of cecal microbiota composition on chicken growth performance and its potential mechanism was investigated.The genera Lachnoclostridium, Alistipes, Negativibacillus, Sellimonas, Sellimonas, and Ruminococcus torques group had beneficial effects on the growth performance of chickens, while Phascolarctobacterium correlated with low growth performance.A correlation analysis revealed links between specific gut microbiota and serum metabolites.In conclusion, certain cecal microbiotas might increase the market weight of chickens by promoting the utilization of pantothenic acid and menadione.
The results of this study provide new insights into the role of gut microbiota in regulating the growth performance of chickens and also lay the foundation for the subsequent development of chicken growth-promoting probiotics or prebiotic-related products.However, there are some limitations that should be addressed or avoided in subsequent research.Firstly, this research was a small sample, single-center, cross-sectional study.Secondly, only the cecal microbiota and serum metabolites at the time of slaughter were analyzed, and samples were not collected at different growth stages.Therefore, the causal mechanism between gut microbiota and growth traits needs to be further studied through larger sample sizes, using a multi-center design and applying the innovative research techniques of integrated omics technology.

Animals 2023 , 24 Figure 1 .
Figure 1.Experimental flow chart.The experimental cohort comprises 49 healthy Guizhou yellow chickens; 8 roosters and 8 hens that possessed the highest body weights in the group constituted the high-market-weight group (HC, n = 16), and a similar group was selected for the low-marketweight group (LC, n = 16).Cecal samples were collected and subjected to 16S rRNA sequencing to infer microbial profiles.Concurrent blood samples were collected to perform untargeted metabolomics detection.Cecal and serum metabolites were identified by statistical analysis.

Animals 2023 , 24 Figure 3 .
Figure 3. Differences of cecal microbiota diversity between high-and low-market-weight chickens.Comparison of α-diversity in cecal microbiota between HC and LC using the (A) Shannon and (B) Simpson indices.(C) PCoA of cecal microbiota from HC and LC.HC, high-market-weight chicken group (n = 16).LC, low-market-weight chicken group (n = 16).

Figure 3 .
Figure 3. Differences of cecal microbiota diversity between high-and low-market-weight chickens.Comparison of α-diversity in cecal microbiota between HC and LC using the (A) Shannon and (B) Simpson indices.(C) PCoA of cecal microbiota from HC and LC.HC, high-market-weight chicken group (n = 16).LC, low-market-weight chicken group (n = 16).

Figure 4 .
Figure 4. Differences of cecal microbiota between high-and low-market-weight chickens.Comparison of cecal microbiota composition between HC/LC at the (A) phylum and (B) genus levels (relative abundance > 0.05%).Cecal microbiota showing different abundances at the (C) phylum and (D) genus

Figure 5 .
Figure 5. Co-abundance groups (CAG) of cecal microbiota associated with high-and low-marketweight chickens.(A) Association of CAGs with HC or LC.CAG6 was enriched in the cecal microbiota of LC, and CAG9, CAG16, and CAG17 were enriched in the cecum of HC. (B) Comparison of mean relative abundance in four differential CAGs between HC and LC.HC, high-market-weight chicken group (n = 16).LC, low-market-weight chicken group (n = 16).

Figure 5 .
Figure 5. Co-abundance groups (CAG) of cecal microbiota associated with high-and low-marketweight chickens.(A) Association of CAGs with HC or LC.CAG6 was enriched in the cecal microbiota of LC, and CAG9, CAG16, and CAG17 were enriched in the cecum of HC. (B) Comparison of mean relative abundance in four differential CAGs between HC and LC.HC, high-market-weight chicken group (n = 16).LC, low-market-weight chicken group (n = 16).

Animals 2023 , 24 Figure 6 .
Figure 6.Identification of the serum metabolites enriched in high-and low-market-weight chickens.(A) Partial least squares discriminant analysis (PLS-DA) of untargeted serum metabolome data from HC and LC.The pink and blue markers represent HC and LC, respectively.(B) Variable importance in projection (VIP > 1) scores for the top serum metabolites contributing to variation in metabolic profiles of HC and LC (Only showed the top 30, the annotation of metabolites are showed in table

Figure 6 .
Figure 6.Identification of the serum metabolites enriched in high-and low-market-weight chickens.(A) Partial least squares discriminant analysis (PLS-DA) of untargeted serum metabolome data from HC and LC.The pink and blue markers represent HC and LC, respectively.(B) Variable importance in projection (VIP > 1) scores for the top serum metabolites contributing to variation in metabolic profiles of HC and LC (Only showed the top 30, the annotation of metabolites are showed in Table 4).Metabolic pathways enriched in (C) HC and (D) LC.HC, high-market-weight chicken group (n = 16).LC, low-market-weight chicken group (n = 16).

Table 2 .
Body weights of Guizhou yellow chickens at the age of 18 weeks.

Table 3 .
The ASVs showing different abundances between high-and high-market-weight chicken groups.

Table 4 .
Annotation of metabolites enriched in high-and low-market-weight-chicken groups.